LEAF COLOUR CHART-GUIDED PRESCRIPTION MAPPING FOR UNMANNED VEHICLE-ASSISTED NITROGEN DISPENSATION IN RICE (ORYZA SATIVA L.)
Rathinavel S, R. Kavitha*, Surendrakumar A, Balaji Kannan1, R. Raja2, S. D. Sivakumar3 and M. K. Kalarani4,
Department of Farm Machinery & Power Engineering, Tamil Nadu Agricultural University, Coimbatore – 641003,
1Department of Soil & Water Conservation Engineering, Tamil Nadu Agricultural University, Coimbatore – 641003,
2ICAR – Central Institute of Cotton Research – Regional Station, Coimbatore – 641003
3Institute of Agriculture, Tamil Nadu Agricultural University, Kumulur, Trichy – 621712,
4 Directorate of Crop Management, Tamil Nadu Agricultural University, Coimbatore – 641003,
*Corresponding author: kavitha@tnau.ac.in
ABSTRACT
Nitrogen management through a leaf colour chart proved to be a cost-effective strategy for small-scale farmers. Identifying a research gap in the realm of variable rate technology, the integration of leaf colour chart readings into precision fertilizer application machinery became the focus. Consequently, a study was initiated to utilize the leaf colour chart data in creating digital prescription maps for unmanned aerial or ground vehicles. Two experimental rice fields were established and assessed for nitrogen deficiency using the leaf colour chart, paired with GPS coordinates. The collected data underwent processing in Google Earth Pro and ArcMap softwares for spatial interpolation of observed data to generate digital prescription maps through predicted data. The Karl Pearson correlation coefficient and Goodness of Prediction (G) of the predicted values with observed values were 0.93 and 0.85. The statistical analysis also confirms (p and f value) the non-existence of significant differences between observed and predicted data. The developed maps were extracted with set of geopositioning and corresponding prescription categories as input data to the precision machinery. This research highlights the potential application of leaf colour chart data and geographical information system tools in unmanned precision fertilizer application machinery.
Keywords: GIS; Leaf colour; Nitrogen management; Precision farming; Rice nutrition.
INTRODUCTION
Globally, rice (Oryza sativa L.) plays a crucial role as a primary agricultural crop in eighty-nine countries, serving as the primary nutritional source for half of the world's population (Adhikari et al., 2023). Nitrogen (N) stands out as the most essential nutrient for plants due to its pivotal role in various plant activities, often acting as a limiting factor in both plant growth and crop production (Dubey et al., 2021). The nitrogen use efficiency (NUE) in rice is currently estimated at only 30 to 40%, with approximately one-third of the applied N being lost through various pathways (Abrol et al., 2007). Nitrogen is lost by various modes such as 10-40% denitrification (Sahu and Samant, 2006), 15-45% nitrification-denitrification process in rice soils (Russo, 1996), as negligible to 50% ammonia vapours (Keller and Men-gel, 1986), 23.9% volatilization from rice fields (Yu et al., 2013).
Numerous researchers aim to enhance NUE by minimizing losses and increasing N uptake in rice crops. To enhance NUE, a common effective approach involves optimizing N utilization through variable rate technology. This optimization relies on high spatial resolution sampling of crop parameters, achievable through either proximal or remote sensing technologies (Mouazen et al., 2020). However, the measurement techniques employed must be rapid, cost-effective, and convenient (Kodaira and Shibusawa, 2013). Various decision support tools, such as Soil Plant Analysis Development (SPAD), (Ban et al., 2022), GreenSeeker, Leaf Colour Chart (LCC), Chlorophyll meter, Yara-N sensor, Urea Super Granules (USG), and split application methods, have demonstrated their effectiveness in N management, contributing to the increased NUE in paddy fields (Lee, 2021). Various sophisticated data acquisition systems, including multispectral imaging and simultaneous or subsequent image processing, can autonomously provide digital geotagged data like NDVI, NDRE, and CCI, enabling drones to distribute N. In contrast, the least expensive option, LCC, is a manual tool that farmers can easily utilize without extensive skill requirements (Leghari et al., 2016). With seven colour shades of greenness, LCC was developed by Furuya (1987) followed by International Rice Research Institute with five or six shades and there are lot many developments done in LCC like Mirror made LCC in Pakistan (Leghari et al.,2016), wheat and maize LCC in India (Singh et al., 2014).
A study revealed that wheat crops managed with LCC-based N practices show a significantly higher profit of nearly Rs.7,000 ha-1 (Mondal and Mitra, 2022). Similarly employing LCC for N management in paddy during active tillering and panicle initiation stages can lead to a more than 9.5% increase in paddy grain yield (Gudadhe and Shah, 2022). Additionally, it helps reduce NO3- leaching, NO2- emissions and enhances both gross and net monetary returns. Notably, omitting basal application resulted in a savings of 10 kg N ha-1 through LCC-based N management practices (Lal et al., 2021). The LCC developed by Punjab Agricultural University and the one by the International Rice Research Institute (IRRI), Philippines, can effectively estimate basmati rice grain yield in-season. Due to their economic feasibility, LCCs, in comparison to chlorophyll meters and canopy reflectance sensors, are viable options for small and marginal farmers in developing countries (Singh et al., 2022).
Precision agricultural machinery (Nath, 2024; Tey et al., 2024) is a wide sector with numerous categories including precise fertilizer applicating machinery. The major two categories of fertilizer application machinery are real-time and map-based. Real-time application machinery may not rely on GPS, as they make use of encoders and onboard sensors mounted on the machinery. The map-based machinery utilizes GPS (Qi et al., 2020; Rathinavel et al., 2023a), GLONASS (GLObal NAvigation Satellite System), BeiDou, EGNOS, SBAS, and similar radio navigation tools for navigation and spatial information. The input data is essentially to be georeferenced, for precise placement of fertilizer. The machinery might be tractor-based (Mirzakhaninafchi et al., 2021), UAV-mounted (Song et al., 2021; Su et al., 2022) or autonomous, also the state of fertilizer might besolid (Zhang et al., 2020) or liquid (Zhou et al., 2023). Shortly, autonomous and robotic machinery was about to boom, the data supplied for those unmanned precision machineries are supposed to be digitized and detailed, to achieve the maximum possible precision.
The existing practice of using LCC is still in manual mode, in which farmer observes a portion of land through several observations and arrives at a single dose of N application based on the average of observed values. As discussed earlier, we need digital spatial data of LCC value throughout the field, for applying LCC in precision machinery. Hence, against the backdrop of enhancing NUE through a straightforward approach, this study aimed to digitize LCC data with geotags and develop digital N management maps. The objective was to facilitate the prescription of precision fertilizer application machinery for rice crops.
MATERIALS AND METHODS
The outline of the study representing workflow is presented in Fig.1. To carry out the study a field experiment was done followed by data acquisition, software processing, and analysis, explained as follows.

Fig. 1. Workflow of the study
Study area: To conduct the designated research, two experimental rice fields were set up at Tamil Nadu Agricultural University in Kumulur, Trichy during the year 2023. The average rainfall of the Kumulur campus was 862 mm (Rathinavel et al., 2020), and located 72 m above mean sea level. Temperature ranges between 25 to 32° C and mean was 28.8° C.
Experiment details: The experimented field is shown in Fig. 2. The chosen rice variety, TRY 3, was cultivated through the transplanting method (Mechanical transplanting with 8-row, riding type Yanmar YP8DN transplanter). A spacing of 30x20 cm was maintained through a mechanical transplanter. Being medium duration crop variety, the fertilizer application plan involved four splits of N, encompassing the recommended doses of, phosphorus, and potassium in a single basal dose. Hand weeding was done twice and no significant weed effect or pest infestations were found. Irrigation was followed for the required times regularly ensuring adequate moisture. The population was consistently maintained, and other variables were kept constant throughout the study. For the TRY 3 variety, the critical LCC value was 4, as per the recommendation from TNAU.
 
Fig. 2. Experimental rice field
 
Fig. 3. LCC and data recording
LCC and data collection: LCC (Fig. 3) threshold from 2 to 5 was applied in conjunction with a hand-held Garmin unit (Model: Garmin GPS etrex 10) for the collection of GPS data (latitude and longitude) in the two experimented fields at regularly spaced (3m x 3m) observation points. At each observation point 10 plants were selected and 6 leaves for each plant were selected randomly. Fully opened, young leaves (fully expanded middle part) from non-diseased plants were specifically chosen for recording observations, during the morning hours from 8:00 AM to 10:00 AM under bright sunshine, with a single person shielding the leaves and chart with body. The critical stages of active tillering and panicle initiation, identified as the most crucial periods (Gudadhe and Shah, 2022), were observed in two experimented rice fields. The boundaries of field I and II derived from Google Earth Pro Environment was illustrated in Fig.4. The boundary files were extracted as KML (Keyhole Markup Language) layer files and saved for further processing. Subsequently, the data, including GPS coordinates and average LCC values (threshold 2 to 5), were plotted using Microsoft Excel.
 
(a) (b)
Fig. 4. Boundaries of field I (a) and field II (b) in Google Earth Pro Environment
Software and input data processing: Similar to the approach taken in soil sampling and mapping, interpolation techniques were adopted as suggested in previous studies (Khan et al., 2021; Kumar et al., 2021). Several researchers have utilized this methodology for investigating spatial variability (Reza et al., 2019; Jin et al., 2021; Malik et al., 2022; Abdu et al., 2023).
The interpolation works based on the principle that unsampled location values were estimated (equation 1) from the neighbouring sampled location. Based on the distance of sampled location, the weights (equation 2) are considered. The Inverse Distance Weighting (IDW) weights are typically proportional to the inverse of the squared distance between the predicted value and the observed value, and they total to 1.
(1)



(2)

The above process was done from a Geographical Information System (GIS) based software. The data recorded in an MS Excel sheet was imported into appositional point data file into ArcMap (version 10.7.1) with X (latitude), Y(longitude), and Z (LCC data). The boundary KML file developed from Google Earth Pro was imported and laid with the point data. The IDW spatial analyst tool was employed to interpolate the data within the defined field boundary. The interpolation process results in a generation of maps. Resulting maps were overlaid with a grid of cells using fishnet and sample tools. The average value for each grid was calculated and then classified to provide recommendations for fertilizer dosage. To assess the accuracy of the developed maps, test values recorded away from the observation points were used.
Data Analysis: Descriptive statistics were analysed for the observed LCC data for both fields and stages. The predicted data extracted from the prescription maps developed was correlated with test data recorded using the Karl Pearson correlation coefficient. Mean Absolute Error (Equation 3), Mean Square Error (Equation 4), Root Mean Square Error (Equation 5) and Coefficient of Determination (Equation 6) ranging from 0 to 1 was determined from the predicted and test data using the following formulae (Yao et al., 2013).
(3)
(4)
(5)
(6)
MAE – Mean Absolute Error
MSE – Mean Square Error
RMSE – Root Mean Square Error
G – Goodness of Prediction
n – Number of samples
x – Observed value
y – Predicted value
A – Mean of Observed values
Using R software, ANOVA was carried out for each set of experiment data. p-value and F-value were discussed.
RESULTS
Digital N management maps and recommendations were generated based on the recorded LCC values observed in two rice fields. The statistical analysis of the observed LCC data was presented in Table 1. The LCC data from the field I indicate a platykurtic distribution with fewer extreme values in both stages. Conversely, field II exhibits leptokurtic and platykurtic kurtosis in stages 1 and 2, respectively, suggesting slight extremes in stage 1. Skewness analysis reveals a rightward skew in both stages of field I and stage 1 of field II, while field II stage 2 exhibits a leftward skew. Overall, the standard deviation and variance indicate a relatively small variance, as the values are not significantly distant from each other.
Table 1. Basic statistics of observed LCC values at different stages of rice growth.
Parameters
|
Field I
|
Field II
|
Active Tillering Stage
|
Panicle Initiation Stage
|
Active Tillering Stage
|
Panicle Initiation Stage
|
Mean
|
3.00
|
3.20
|
2.90
|
3.30
|
Median
|
3.00
|
3.00
|
3.00
|
3.50
|
Mode
|
3.50
|
3.00
|
3.00
|
3.50
|
Range
|
2.50-3.50
|
2.50-4.00
|
2.50-3.50
|
2.50-4.00
|
Standard Deviation
|
0.37
|
0.42
|
0.26
|
0.41
|
Variance
|
0.14
|
0.18
|
0.06
|
0.17
|
Skewness
|
0.13
|
0.64
|
0.32
|
-0.67
|
Kurtosis
|
-1.40
|
-0.38
|
0.17
|
-0.11
|
The Fig. 5 depicts a strong positive correlation between all 4 (both fields and both stages) datasets of observed test value and predicted value (derived from interpolated maps). Table 2, Fig. 6 and Fig. 7 represents the relationship between observed and predicted data with other statistical data. There is no significant difference exists between observed and predicted values in all 4 sets of data.

Fig. 5. Correlation between Observed and Predicted data sets
Table 2. Evaluation of interpolated LCC map.
Parameter
|
Field I
|
Field II
|
Mean
|
Active Tillering Stage
|
Panicle Initiation Stage
|
Active Tillering Stage
|
Panicle Initiation Stage
|
Mean Absolute Error
|
0.11
|
0.068
|
0.065
|
0.082
|
0.081
|
Mean Square Error
|
0.024
|
0.017
|
0.01
|
0.026
|
0.01925
|
Root Mean Square Error
|
0.15
|
0.13
|
0.10
|
0.16
|
0.135
|
Goodness of Prediction
|
0.85
|
0.86
|
0.95
|
0.76
|
0.855
|
p-value
|
0.81
|
0.75
|
0.95
|
0.94
|
|
F-value
|
4.09
|
4.11
|
4.09
|
4.09
|
|

Fig. 6. Histogram for Active tillering stage

Fig. 7. Histogram for Panicle initiation stage
A digital map based on LCC for N management was created, as illustrated in Fig. 8. Previous studies by Wollenhaupt et al. (1994), Chan et al. (2004), Hedley et al. (2012), and Mezera et al. (2018) have classified zones differently based on specific field requirements. The map was divided into three zones according to the observed LCC values: Zone I (2.5 to 3.0), Zone II (3.0 to 3.5), and Zone III (3.5 to 4). In the active tillering stage of Field I, classes I, II, and III covered approximately 24%, 42%, and 38%, respectively. In the panicle initiation stage, these percentages changed to 38%, 33.5%, and 28.5%. The recommendations generated (Table 3) were validated through multiple replications within each class, comparing them with the statistics of LCC data. This validation affirms that Field I does not exhibit extreme values.
 
(a) (b)
Fig. 8. LCC-based digital map on N management for the field I at (a) active tillering and (b) panicle initiation stages
Table 3. Recommended class of fertilizer dosage in field I.
GPS Coordinates
|
78.82799⁰E
|
78.82811⁰E
|
78.82824⁰E
|
10.93169⁰N
|
I
|
I
|
II
|
I
|
II
|
I
|
10.93161⁰N
|
I
|
I
|
I
|
I
|
II
|
I
|
10.93154⁰N
|
I
|
II
|
I
|
I
|
II
|
I
|
10.93146⁰N
|
II
|
II
|
II
|
II
|
II
|
II
|
10.93139⁰N
|
II
|
II
|
II
|
II
|
III
|
II
|
10.93132⁰N
|
III
|
III
|
II
|
III
|
III
|
III
|
10.93124⁰N
|
III
|
III
|
III
|
III
|
III
|
III
|
In the case of field II, three categories of zones were identified, denoted as I (2.5 to 3.0), II (3.0 to 3.5), and III (3.5 to 4). During the active tilleringstage, these classes covered approximately 20%, 80%, and 0% of the area, respectively. In the panicle initiationstage, the distribution shifted, with classes I, II, and III covering around 15%, 65%, and 20% of the area in field II. The resulting digital map (Fig. 9.) exhibited a subtle roughness consistent with the statistics from the observed data. The recommendation class of fertilizer dose was tabulated in Table 4.

(a)

(b)
Fig. 9.LCC-based digital map on N management for field II ((a) active tillering and (b) panicle initiation stages)
Table 4. Recommended class of fertilizer dosage in field II.
GPS Coordinates
|
78.83066⁰E
|
78.83078⁰E
|
78.8309⁰E
|
78.83102⁰E
|
78.83114⁰E
|
10.93222⁰N
|
II
|
III
|
II
|
II
|
II
|
II
|
II
|
II
|
II
|
II
|
10.93214⁰N
|
II
|
III
|
II
|
II
|
II
|
II
|
II
|
II
|
II
|
II
|
10.93205⁰N
|
II
|
II
|
I
|
II
|
I
|
I
|
II
|
II
|
10.93197⁰N
|
II
|
III
|
I
|
I
|
II
|
II
|
10.93189⁰N
|
II
|
III
|
I
|
I
|
II
|
II
|

Fig. 10. Area distribution of three different classes on either fields or stages
DISCUSSION
Despite noting an increase from the active tillering to the panicle initiation stage, a majority of the field areas experienced N deficiency, categorized with varying intensity in both stages and plots. From Fig. 10.it is observed that majority of the area falls into class II, followed by class I, and last class III. From this area distribution of different classes, a 12.6% of N can be saved if the adjusted N application was set to 75-100% in comparison with blind application of 100% N to the entire field. This saves 32.5 kg of urea, which in turn saves cost, energy, labor, machinery wear, etc. Though the savings appears to be less in this experiment, situations with wider variations may have potential N savings.
The mean correlation coefficient was 0.93 and a goodness of fit value of 0.85. This implies a strong feasibility of the proposed technique for precision machinery applications. Similarly, Morgan et al. (2017)proved the use of the IDW technique in the agricultural sector and obtained a coefficient of correlation of more than 0.80.
Shukla et al. (2004) reported a correlation ranging from 0.84 to 0.91 between LCC and chlorophyll content (SPAD observations) across all rice genotypes. This suggests that the developed map from LCC data could theoretically demonstrate a positive relationship of 78.5% to 85% with chlorophyll content. Different varieties of rice had significant effect on yield for LCC based N management (Bhat et al., 2022).LCC use saved 10 to 40 kg N ha–1 (Bhat, 2014). Hence the efficiency of the technique depends on the efficiency of LCC towards the variety also. Since, GPS is satellite-based navigation tool (Chaudhary et al., 2024), the accuracy of positioning and navigation is crucial for a precision machinery (Nijak et al., 2024). Hence the GPS is main affecting factor which is influencing the data acquisition and also the applicating machine efficiency.
The positive relationship between the classification data and observed data statistics, as previously discussed, validates the accuracy of the developed map. With the recommended class and geotags, unmanned vehicles, whether aerial or ground-based, can efficiently dispense fertilizer using mechanisms such as a fluted roller or a variable flow liquid spray system. This approach is particularly viable for small farms where manual data acquisition is not cumbersome, showcasing the potential for precision farming in the predominant small-scale agriculture sector in India.
Table 5. Cost of various precision fertilizer management tools
S. No
|
Tool
|
Cost (Rs.)
|
1
|
GreenSeeker
|
1,50,000/-
|
2
|
Opti-Science Chlorophyll meter
|
1,00,000 – 1,50,000/-
|
3
|
SPAD Chlorophyll meter
|
1,00,000 – 1,50,000/-
|
4
|
Multispectral camera (Drone)
|
8,00,000 – 20,50,000/-
|
5
|
LCC
|
30/-
|
6
|
Spectrophotometer
|
50,000 – 5,00,000/-
|
The advantage of the proposed technique lies in the cheaper cost involvement in comparison with sophisticated technologies like SPAD meter, chlorophyll meter and multispectral imaging which are regarded as costlier (Table 5) in terms of data acquisition (Andrianto et al., 2020; Kamarianakis and Panagiotakis, 2023). Further, the use of LCC by farmers was an existing practice and familiar to extend towards them. The technology holds good for an entrepreneur to undertake a precision agriculture custom hiring or project.
The limitation of the technology lies in the requirement of a skilled assistance for map generation and operating precision machinery based on the map. Accuracy and precision of results may not be on par with drone imaging and other sensor based techniques. Low-Cost Chlorophyll Content Meter developed by Kamarianakis and Panagiotakis (2023) had R2 values of 0.95 to 0.99 found between their low cost chlorophyll content meter and SPAD 502 in different crops. GPS accuracy and boundary marking must be under utmost care for the successful results. Another fact is LCC is applicable only to the top-dressed N application and not suitable for basal N application (Rao and Das, 2023).
In future, with 5G enabled network facilities (Rathinavel et al., 2023b), smart phone adoption by farmers, robust software application etc., will be easy to eliminate the limitations discussed above and can be a small scale technology for small, marginal farmers and custom hiring centres.
Conclusion: By addressing the research gap in variable rate technology, the study successfully generated digital N management maps for two experimental rice fields, achieving a significant correlation of more than 93% and 85% of goodness of fit models with ground truth test values. The utilization of LCC data and GIS tools presents a cost-effective and practical strategy for enhancing NUE in small-scale farming, showcasing the potential for widespread application in precision farming for the benefit of farmers in developing countries.
REFERENCES
- Abdu, A., F. Laekemariam, G. Gidago,A. Kebede and L. Getaneh (2023). Variability analysis of soil properties, mapping, and crop test responses in Southern Ethiopia. Heliyon. 9(3): e14013. doi: 10.1016/j.heliyon.2023.e14013.
- Abrol, Y. P., R. Pandey, N. Raghuram and A. Ahmad (2012). Nitrogen cycle sustainability and sustainable technologies for nitrogen fertilizer and energy management. J. Indian Ins. Sci., 92(1): 17-36.
- Adhikari S., S.K. Sah, S. Gautam, D. Subedi, and S. Katel (2023). Leaf colour chart based nitrogen management in spring rice. Fundamental App. Agric. 8(1&2): 368–376. doi: 10.5455/faa.99074.
- Andrianto, H., S. Suhardi and A. Faizal (2020). Performance evaluation of low-cost IoT based chlorophyll meter. Bullet. Electric. Engg. Informatics. 9(3): 956-963. doi: 10.11591/eei.v9i3.2014.
- Ban, S., W. Liu, M. Tian, Q. Wang, T. Yuan, Q. Chang and L. Li (2022). Rice Leaf chlorophyll content estimation using uav-based spectral images in different regions. Agron. 12: 2832. doi: 10.3390/agronomy12112832.
- Bhat, T.A., R.H. Kanth, B. Jan, A. Nazir, S.A. Ahanger, M.S. Mir, N.B. Naikoo, S. Fayaz, K.A. Dar and A. Gul (2022). Real-time nitrogen application of rice varieties based on leaf colour chart under system of rice intensification in temperate climate. Agron. 12(9): 2229. doi: 10.3390/agronomy12092229.
- Bhat, T.A. (2014). Real time nitrogen management in rice genotypes (Oryza sativa) based on Leaf Colour Chart (LCC), Doctoral thesis (unpublished), Sher-e-Kashmir Univ. Agric. Sci. Tech., Kashmir, India.
- Chan, C.W., J.K. Schueller, W.M. Miller, J.D. Whitney and J.A. Cornell (2004). Error Sources Affecting Variable Rate Application of Nitrogen Fertilizer. Prec. Agric. 5: 601–616. doi.org/10.1007/s11119-004-6345-2.
- Chaudhary, R.R., S. Jain and S. Gupta (2024). A critical review on hybrid framework for precise farming with application of Machine Learning (ML) and Internet of Things (IoT). J. Integrated Sci. Tech. 12(2): 730-730.
- Dubey, R.S., R.K. Srivastava and M. Pessarakli (2021). Physiological mechanisms of nitrogen absorption and assimilation in plants under stressful conditions. In Handbook of plant and crop physiology, CRC Press, pp 579-616.
- Furuya, S. (1987). Growth diagnosis of rice plants by means of leaf color. Jpn. Agr. Res. Q. 20: 147-153.
- Gudadhe, N. and K.A. Shah (2022). Real-time nitrogen management in rice through leaf colour chart, ICAR, AGRIS. FAO. Int. Sys. Agric. Sci. Tech. 72(2): 34-37.
- , Carolyn and Y. Ian (2012). Farmer uptake of variable rate irrigation technologies in New Zealand, 10th Int. Conf. Prec. Agric, Denver, Colorado, USA.
- Jin, Z., L. Zhang, L.V., J and X. Sun (2021). The application of geostatistical analysis and receptor model for the spatial distribution and sources of potentially toxic elements in soils. Env. Geochem. Health. 43: 407-421. doi: 10.1007/s10653-020-00729-6.
- Kamarianakis, Z. and S. Panagiotakis (2023). Design and implementation of a low-cost chlorophyll content meter. Sensors. 23(5): 2699. doi: 10.3390/s23052699.
- Keller G.D. and D.B. Mengel (1986). Ammonia volatilization from nitrogen fertilizers surface applied to no-till Soil Sci. Society America J. 50:1060–1063. doi: 10.2136/sssaj1986.03615995005000040045x.
- Khan, M.Z., M.R. Islam, A.B.A. Salam and T. Ray (2021). Spatial variability and geostatistical analysis of soil properties in the diversified cropping regions of Bangladesh using geographic information system techniques. App. Env. Soil Sci. doi: 10.1155/2021/6639180.
- Kodaira, M. and S. Shibusawa (2013). Using a mobile real-time soil visible-near infrared sensor for high resolution soil property mapping. Geoderma. 199: 64–79. doi: 10.1016/j.geoderma.2012.09.007.
- Kumar, P., P. Kumar, M. Sharma, N.P. Butail and A.K. Shukla (2021). Quantification of the soil organic carbon and major nutrients using geostatisticalapproach for Lahaul valley, cold arid region of trans-himalaya. Adv. Carbon Capture Utilization. 235-247. doi: 10.1007/978-981-16-0638-0_10.
- Lal, B S., K. Lamsal, D.N. Yogi and G.R. Chaudhary (2021). Real time nitrogen management in rice using leaf colour chart under rainfed condition of western hills of Nepal. Int. J. Env. Agric. Res. 7(7): 9-46.
- Lee, S. (2021). Recent advances on nitrogen use efficiency in rice. Agron. 11(4): 753. doi: 10.3390/agronomy11040753.
- Leghari, S.J., U.A. Leghari, M. Buriro, G.M. Laghari, F.A. Soomro, M.A. Khaskheli and S.S. Hussain (2016). Modern leaf colour chart successfully prepared and used in crop production of Sindh, Pakistan. European Academic Res. J. 4(2): 900-916.
- Malik, S., T. Bhowmik, U. Mishra and N. Paul (2022). Mapping and prediction of soil organic carbon by an advanced geostatistical technique using remote sensing and terrain data. Geocarto Int. 37(8): 2198-2214. doi: 10.1080/10106049.2020.1815864.
- Mezera, J., L. Vojtech, J. Elbl, A. Kintl and V. Smutny (2018). Evaluation of variable rate application of fertilizers by proximal crop sensing and yield mapping, 19th Multidisciplinary Sci. GeoConf. SGEM 2019, Vienna, Austria, 978-619-7408-80-5: 831-838.
- Mirzakhaninafchi, H., M. Singh, V. Bector, O.P. Gupta and R. Singh (2021). Design and development of a variable rate applicator for real-time application of fertilizer. Sustainability. 13(16): 8694. doi: 10.3390/su13168694.
- Mondal, T. and B. Mitra (2022). Leaf colour chart based nitrogen management in wheat under zero till condition of West Bengal. Indian Farming. 71(12).
- Morgan, R. S., M. Abd El-Hady, I.S. Rahim, J. Silva and S. Ribeiro (2017). Evaluation of various interpolation techniques for estimation of selected soil properties. Geomate J. 13(38): 23-30. doi: 10.21660/2017.38.606721.
- Mouazen, A.M., T. Alexandridis, H. Buddenbaum, Y. Cohen, D. Moshou, D. Mulla, S. Nawar and K.A. Sudduth (2020). Monitoring agricultural internet of things and decision support for precision smart farming. elsevier. pp 35–138.
- Nath, S. (2024). A vision of precision agriculture: Balance between agricultural sustainability and environmental stewardship. Agron. J. 116(3): 1126-1143. doi: 10.1002/agj2.21405.
- Nijak, M., P. Skrzypczyński, K. Ćwian, M. Zawada, S. Szymczyk and J. Wojciechowski (2024). On the importance of precise positioning in robotised agriculture. Remote Sensing. 16(6): 985. doi: 10.3390/rs16060985.
- Qi, J., X. Tian, Y. Li, X. Fan, H. Yuan, J. Zhao and H. Jia (2020). Design and experiment of a subsoiling variable rate fertilization machine. Int. J. Agric. Biological Engg. 13(4): 118-124. doi: 10.25165/j.ijabe.20201304.5757.
- Rao, K.G. and N.K. Das (2023). Leaf color chart (LCC) a reliable tool for nitrogen management in (Dry seeded rice and transplanted rice): A review. J. Pharm. Innov. 12: 3974-3979.
- Rathinavel, S., B. Kannan, R. Kavitha and J. Ramachandran (2023a). Application of GIS in farm mechanization. Just Agric. 3(6): 9-12.
- Rathinavel, S., R. Kavitha, J. Gitanjali and R. Saiprasanth (2023b). Role of 5G technology in enhancing agricultural mechanization. IOP Conf. Series: Earth Environ. Sci. 1258(1): 012010. doi: 10.1088/1755-1315/1258/1/012010.
- Rathinavel, S., A. Mani, Jithendrakumar, P. Anbarasan, P. Soundharya and S. Soundharya (2020). Drought assessment in Tiruchirapalli district using varied time scales. Indian J. Ecology. 47(2): 295-298.
- Reza, S.K., D. Dutta, S. Bandyopadhyay and S.K. Singh (2019). Spatial variability analysis of soil properties of Tinsukia District, Assam. India. Agric. Res. 8: 231-238. doi: 10.1007/s40003-018-0365-z.
- Russo S. (1996). Rice yield as affected by the split method of N application and nitrification in-hibitor DCD. Cahiers Options Mediterraneennes. 15: 43–53.
- Sahu, S.K. and P.K. Samant (2006). Nitrogen loss from rice soils in Orissa, Orissa Review, India. pp 34–36.
- Shukla, A.K., V.K. Singh, B.S. Dwivedi, S.K. Sharma and P.S. Pandey (2004). Calibrating the leaf colour chart for nitrogen management in different genotypes of rice and wheat in a systems perspective. Agron. J. 96(6): 1606-1621. doi: 2134/agronj2004.1606.
- Singh, V., B. Singh, H.S. Thind, Y. Singh, R.K. Gupta, S. Singh and V. Balasubramanian (2014). Evaluation of leaf colour chart for need-based nitrogen management in rice, maize and wheat in north-western India. J. Res. Punjab Agric. Univ. 51(3-4): 239-245.
- Singh, V., M. Singh and B. Singh (2022). Spectral indices measured with proximal sensing using canopy reflectance sensor, chlorophyll meter and leaf colour chart for in-season grain yield prediction of basmati rice. Pedosphere. 32(6): 812-822. doi: 10.1016/j.pedsph.2022.06.015.
- Song, C., Z. Zhou, Y. Zang, L. Zhao, W. Yang, X. Luo and Q. Zhu (2021). Variable-rate control system for UAV-based granular fertilizer spreader. Comput. Electronics Agric. 180: 105832. doi: 10.1016/j.compag.2020.105832.
- Su, D., W. Yao, F. Yu, Y. Liu, Z. Zheng, Y. Wang and C. Chen (2022). Single-neuron PID UAV variable fertilizer application control system based on a weighted coefficient learning correction. Agric. 12(7): 1019. doi: 10.3390/agriculture12071019.
- Tey, Y.S., M. Brindal, S.Y. Wong, Ardiansyah, A. Ibragimov and M.R. Yusop (2024). Evolution of precision agricultural technologies: a patent network analysis. Prec. Agric. 25(1): 376-395. doi: 10.1007/s11119-023-10076-y.
- Wollenhaupt, N.C., R.P. Wolkowski and M.K. Clayton (1994). Mapping soil test phosphorus and potassium for variable-rate fertilizer application. J. Prod. Agric. 7(4): 395-404. doi: 10.2134/jpa1994.0441.
- Yao, X., B. Fu, Y. Lü, F. Sun, S. Wang and M. Liu (2013). Comparison of four spatial interpolation methods for estimating soil moisture in a complex terrain catchment. PloS one. 8(1): e54660. doi: 10.1371/journal.pone.0054660.
- Yu Q., J. Ye, S. Yang, J. Fu, J. Ma, W. Sun, L. Jian and Q. Wang (2013). Effects of nitrogen application level on rice nutrient uptake and ammonia volatilization. Rice Sci. 20: 139–147. doi: 10.1016/S1672-6308(13)60117-1.
- Zhang, Z., C. Chen, H. Li and H. Xia (2020). Design, development, and performance evaluation of a fertigation device for distributing solid fertilizer. Water. 12(9): 2621. doi: 10.3390/w12092621.
- Zhou, W., T. An, J. Wang, Q. Fu, N. Wen, X. Sun and Z. Liu (2023). Design and experiment of a targeted variable fertilization control system for deep application of liquid fertilizer. Agron. 13(7): 1687. doi: 10.3390/agronomy13071687.
|